HPC & AI Performance Engineer

About the role

HPE operates at the forefront of a fast-moving HPC and AI industry. We are looking for driven engineers who thrive on complex technical challenges and consistently deliver high-quality results that strengthen the performance, competitiveness, and success of HPE’s HPC & AI solutions for customers worldwide. This position will support government accounts. Therefore, due to federal export-control regulations, the selected candidate must hold U.S. citizenship, U.S. lawful permanent resident/Green Card status or otherwise have a category of refugee/asylee status enabling them to perform the role without requiring a license under the International Traffic in Arms Regulations (ITAR) or Export Administration Regulations (EAR).

Responsibilities

• Lead HPC and AI benchmarking projects across CPU, GPU, network, memory, and storage platforms.
• Evaluate HPE and competitive HPC and AI architectures using performance models and benchmark data.
• Run and analyze scientific, engineering, large language model (LLM), AI training and inference, storage, and I/O workloads.
• Identify performance bottlenecks and optimize applications, AI frameworks, libraries, compilers, runtimes, and system software.
• Apply parallel computing, GPU acceleration, profiling, and memory and I/O optimization to deliver credible, repeatable results.

Qualifications

• 6+ years of Experience with HPC and AI workloads, including scientific and engineering applications, MLPerf, large language models such as DeepSeek, Kimi K2.6 or K3, and gpt-oss-120b, AI training and inference, storage benchmarks, and performance-profiling tools. Relevant coursework and internship experience will be considered.
• Knowledge of HPC and AI system architecture, including CPUs, GPUs and other accelerators, memory, networking, storage, and software stacks, with the ability to explain their impact on application and benchmark performance.
• Understanding of parallel and distributed programming techniques, including MPI, OpenMP, OpenSHMEM, algorithms, and performance considerations for HPC and AI workloads.
• Ability to lead complex HPC and AI performance projects, work effectively across technical teams, and translate findings into customer-focused recommendations.
• Demonstrated ability to analyze and optimize computational applications and benchmarks on Linux-based HPC and AI systems.
• Experience with HPC and AI software environments, including C, C++, Fortran, Python, Linux scripting, compilers, MPI, MPI-IO, OpenMP, and relevant AI frameworks and libraries.
• Experience with NVIDIA GPUs and AMD Instinct MI-series accelerators, including offloading computational routines using CUDA, HIP, OpenMP target offload, OpenACC, or comparable programming models for HPC and AI workloads.
• Experience using performance-profiling, tracing, and debugging tools on Linux-based HPC and AI systems.
• Ability to interpret benchmark results, identify performance bottlenecks, and recommend improvements for HPC and AI applications and systems.
• Excellent analytical and problem-solving skills.
• Excellent written and verbal communication skills, with the ability to present complex technical findings clearly to engineering teams, customers, and business stakeholders; professional proficiency in English required.
• Master’s degree in computer science, engineering, mathematics, physics, chemistry, environmental science, or a related technical field; PhD preferred.

Benefits

• Health & Wellbeing: comprehensive suite of benefits supporting physical, financial and emotional wellbeing.
• Personal & Professional Development: investment in career with specific programs to help reach career goals, become a knowledge expert or apply skills to another division.
• Unconditional Inclusion: flexible work to manage work and personal needs, bold moves together, force for good.

Pay

United States of America: Annual Salary USD 109,000 - 251,000 in Minnesota & Texas (base salary; variable incentives may also be offered).

Schedule

Hybrid role requiring on average 2 days per week from an HPE office.